The article and associated spreadsheet present a compelling argument for the use of ordinary least-squares regression analysis in calibration and validity studies. This communication was prompted by a spirited debate at the 2004 ACSM Annual Meeting and subsequently via electronic mail. It is worth stating that the analysis and interpretation of the study scenarios presented is one of the most controversial and often impenetrable topics in the measurement and biostatistics literature. The article discusses the most common scenario of two methods of measurement of some underlying quantity, with no replicate measurements. Several authors, including Bland and Altman (1999) and Dunn (2004) have argued that this design is weak as the resulting data are difficult to interpret with any confidence. These authors maintain that repeated measurements of each of the two methods should ideally be incorporated into the analysis, to take account for the influence of varying degrees of measurement error in both methods. This issue has implications for the potential artifactual proportional bias in Bland-Altman plots.